Value betting is staking only when your estimated probability of an outcome is higher than the probability implied by the bookmaker’s odds. The edge is in the price, not in “picking winners” for their own sake.

If decimal odds are 2.00, the book’s implied probability is 1 / 2.00 = 50%. If you estimate the same outcome at 55%, the bet is +EV at that price. If you estimate 45%, it is −EV — even if the selection later wins. For odds formats and conversion, see Sports betting odds explained.

Implied probability and overround

Raw implied probability from decimal odds:

implied_probability = 1 / decimal_odds

Two-way example (illustrative prices only):

Selection Odds Raw implied
Player A 1.80 1 / 1.80 = 55.56%
Player B 2.00 1 / 2.00 = 50.00%
Sum 105.56%

The sum is 105.56%, not 100%. That 5.56 percentage-point gap above 100% is the bookmaker’s overround (margin). Books do not post a fair 100% market; they build in a hold.

A common teaching step is to normalise (divide each raw implied by the sum) so the sides add to 100%:

  • Player A: 55.56% / 105.56% = 52.63%
  • Player B: 50.00% / 105.56% = 47.37%

Call these de-vigged / fair-from-this-market probabilities — not “true” probabilities in any absolute sense. They only strip that book’s overround under a proportional assumption. In practice, value bettors usually take a sharper reference (tight sharp book, exchange, or consensus) as the probability proxy, then look for longer prices elsewhere. Same maths either way: you need a p, then you compare it to the offer.

Expected value (EV)

For a stake S at decimal odds O, with win probability p:

EV = p × S × (O − 1) − (1 − p) × S

which simplifies to:

EV = S × (p × O − 1)

Edge (as a fraction of stake) is p × O − 1. Positive EV means positive edge at that price; it does not mean the next bet wins. See also What is expected value in sports betting?.

Worked example

Keep the de-vigged p = 52.63% for Player A from the table above as your probability proxy. Another book offers 2.20 on Player A. Stake $100:

  • Win: net profit $100 × (2.20 − 1) = $120 with probability 52.63%
  • Lose: −$100 with probability 47.37%

EV = 0.5263 × 120 − 0.4737 × 100 = 63.156 − 47.37 = $15.79

Same result via the compact form:

EV = 100 × (0.5263 × 2.20 − 1) = 100 × 0.15786 = $15.79

So +15.79% of stake in expectation if p = 52.63% is right. If your proxy is wrong, the “edge” is fiction. Illustrative odds only — not a tip and not a live market.

Value betting vs matched betting vs arbitrage

Approach Where the money comes from Outcome risk on the event
Value betting Price longer than your fair p Full — you want that exposure
Matched betting Promo / free-bet terms after a hedge Near zero if stakes are set correctly
Arbitrage Price gap between books (cash both sides) Near zero if both legs are on

Matched betting clears offers; it is finite once welcome/reloads are done. Arbitrage locks a gap but burns soft accounts fast and often needs awkward stake splits. Value betting is the long-run pricing game. Deeper splits: Matched betting explained, Betting arbitrage, Arbitrage bet vs value bet.

Variance and bankroll

+EV does not remove variance. A correct 55% shot still loses 45% of the time. Short samples can be deep red even with a real edge; over a large number of independent bets, results tend toward EV (law of large numbers in sports betting).

Stake size controls ruin risk. Flat stakes are simple. Kelly (and fractional Kelly) sizes from odds, p, and bankroll so you neither overbet nor leave too much growth on the table — see Bankroll management and Kelly. More bets and diversified markets reduce relative variance; they do not turn a −EV process into a +EV one (reducing variance).

Accumulators: multiplying several +EV legs can raise the combined edge on the ticket if each leg is still +EV at the multiplied price and dependence is handled honestly. Hit rate collapses as you add legs, so variance and drawdowns jump. Accas are a staking/structure choice, not a shortcut around bankroll rules. More: Accumulators — upsides and traps.

Trade-offs (plain)

Works when

  • You have a repeatable way to estimate p (model, sharp reference, or disciplined process) better than the soft price on offer
  • You size so losing streaks do not force you out before sample size arrives
  • You accept that soft books will limit or close winning accounts over time

Breaks when

  • “Edge” is just a de-vig of the same soft book you are betting into
  • Stakes are too large relative to bankroll (or to how uncertain p is)
  • You chase results after a cold run instead of process (emotion and betting discipline)

Limits are expected for winners. Practical notes: How to avoid getting limited, Staying under the radar at soft books.

Practical setup

You need accounts that still take meaningful stakes, a probability source, and a stake rule. Scanning many books by hand is slow; tools exist to surface +EV prices against a reference. Trademate Sports is built for that workflow — compare book lists and markets in the bookmaker guide for Trademate users, not invented ROI claims.

Arbitrage can coexist with value work, but arbing soft books usually shortens account life faster than selective +EV. Context: Why many bettors move from arb to value. Live markets add volume and latency risk; they do not change the EV definition (live betting notes).

FAQ

What is a value bet?
A bet where your estimated win probability p satisfies p × odds − 1 > 0. The selection can lose and the bet was still +EV at the time you took it.

Is value betting the same as matched betting?
No. Matched betting hedges a promo so outcome risk is near zero. Value betting keeps the exposure because the price is the edge. See Matched betting.

How is value betting different from arbitrage?
Arbitrage locks both sides across books for a near-certain small profit. Value betting is one-sided (or selectively sided) against a fair p. Comparison: Arbitrage vs value.

Does a +EV bet guarantee profit?
No. EV is an average over many similar bets. One bet has two outcomes: win (odds − 1) × stake or lose −stake.

Why do bookmakers limit value bettors?
Soft books are not priced to lose long-term to sharp money. Sustained +CLV / +EV volume is a signal to cut max stake or close the account. That is a cost of the strategy, not proof the maths failed.

Related

If you want to test the method on real prices, Trademate Sports lists bets where a soft bookmaker's odds sit above a sharp reference line, with the EV shown per bet. Start with Trademate Sports. Edge is not profit: expect variance and account limits.